{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Approach 4: Autoencoder - Boxplot\n",
    "This notebook reads eval results of sets of identical trainings and plots them into a boxplot. The names of the training results can be configured in the cell below."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 3 sessions\n"
     ]
    }
   ],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "\n",
    "import numpy as np\n",
    "from sklearn.neighbors import KernelDensity\n",
    "from glob import glob\n",
    "from sklearn.metrics import roc_curve, auc\n",
    "import matplotlib.pyplot as plt\n",
    "from tqdm import tqdm\n",
    "\n",
    "from py.Dataset import Dataset\n",
    "from py.FileUtils import load\n",
    "from py.PlotUtils import get_percentiles\n",
    "\n",
    "DIR = '/home/kleinsteuber/vscode/ResizedSessions256_NoBackup' # dataset directory\n",
    "CONFIGS_BY_LATENT_FEATURES = [\n",
    "    {\n",
    "        \"label\": \"16\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_16f_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"32\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_32f_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"64\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_64f_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"128\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_128f_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"256\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_256f_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"512\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"1024\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "]\n",
    "\n",
    "CONFIGS_DENOISING = [\n",
    "    {\n",
    "        \"label\": \"0\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"0.100\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_noise0.01_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"0.150\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_noise0.0225_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"0.200\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_noise0.04_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"0.300\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_noise0.09_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"0.400\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_noise0.16_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "]\n",
    "\n",
    "CONFIGS_SPARSE = [\n",
    "    {\n",
    "        \"label\": \"0\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"1e-5\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_sparse1e-5_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"1e-4\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_200e_sparse_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"1e-3\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_sparse1e-3_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"1e-2\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_sparse1e-2_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "]\n",
    "\n",
    "CONFIGS_DENOISING_AND_SPARSE = [\n",
    "    {\n",
    "        \"label\": \"Base\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"σ = 0.1\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_noise0.01_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"σ = 0.15\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_noise0.0225_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"σ = 0.2\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_noise0.04_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"σ = 0.3\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_noise0.09_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"σ = 0.4\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_noise0.16_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"λ = 1e-5\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_sparse1e-5_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"λ = 1e-4\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_200e_sparse_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"λ = 1e-3\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_sparse1e-3_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"label\": \"λ = 1e-2\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_sparse1e-2_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    #{\n",
    "    #    \"label\": \"σ = 0.1, λ = 1e-4\",\n",
    "    #    \"train_names\": [f\"ae2_beaver_01_noise0.01_sparse1e-4_200e_{i}\" for i in range(10)]\n",
    "    #},\n",
    "]\n",
    "\n",
    "CONFIGS_COMPARE_SESSIONS = [\n",
    "    {\n",
    "        \"session\": \"Beaver_01\",\n",
    "        \"label\": \"Beaver_01\",\n",
    "        \"train_names\": [f\"ae2_beaver_01_noise0.01_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"session\": \"Marten_01\",\n",
    "        \"label\": \"Marten_01\",\n",
    "        \"train_names\": [f\"ae2_marten_01_noise0.01_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "    {\n",
    "        \"session\": \"GFox_03\",\n",
    "        \"label\": \"GFox_03\",\n",
    "        \"train_names\": [f\"ae2_gfox_03_noise0.01_200e_{i}\" for i in range(10)]\n",
    "    },\n",
    "]\n",
    "\n",
    "ds = Dataset(DIR)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Calculating the scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_scores(train_names, session = \"Beaver_01\", **kwargs):\n",
    "    kde_scores = {\n",
    "        \"auc\": [],\n",
    "        \"tnr90\": [],\n",
    "        \"tnr95\": [],\n",
    "        \"tnr99\": []\n",
    "    }\n",
    "    loss_scores = {\n",
    "        \"auc\": [],\n",
    "        \"tnr90\": [],\n",
    "        \"tnr95\": [],\n",
    "        \"tnr99\": []\n",
    "    }\n",
    "    # correct name\n",
    "    session = ds.create_session(session).name\n",
    "    for train_name in tqdm(train_names):\n",
    "        try:\n",
    "            lapse_losses, lapse_encodings, lapse_labels, lapse_images = load(f\"./ae_train_NoBackup/{train_name}/eval/{session}_lapse.pickle\")\n",
    "            motion_losses, motion_encodings, motion_labels, motion_images = load(f\"./ae_train_NoBackup/{train_name}/eval/{session}_motion.pickle\")\n",
    "        except:\n",
    "            print(f\"{train_name} is missing, skipping...\")\n",
    "            continue\n",
    "\n",
    "        # # # # # # # # # # # # # # # # # #\n",
    "        # 1. KDE AUC & elimination rates\n",
    "        # # # # # # # # # # # # # # # # # #\n",
    "\n",
    "        kde = KernelDensity(kernel=\"gaussian\", bandwidth=0.01).fit(lapse_encodings)\n",
    "        preds = kde.score_samples(motion_encodings)\n",
    "        y_anom = preds[motion_labels == 1]\n",
    "        y_norm = preds[motion_labels == 0]\n",
    "        # likelihoods need to be inverted (anomalies have lower likelihood)\n",
    "        X = -np.concatenate([y_norm, y_anom]).reshape((-1, 1))\n",
    "        y = np.concatenate([-np.ones_like(y_norm), np.ones_like(y_anom)])\n",
    "\n",
    "        fpr, tpr, thresholds = roc_curve(y, X[:,0])\n",
    "        kde_scores[\"auc\"].append(auc(fpr, tpr))\n",
    "        kde_scores[\"tnr90\"].append(get_percentiles(fpr, tpr, thresholds, percentiles=[0.9], verbose = False)[0])\n",
    "        kde_scores[\"tnr95\"].append(get_percentiles(fpr, tpr, thresholds, percentiles=[0.95], verbose = False)[0])\n",
    "        kde_scores[\"tnr99\"].append(get_percentiles(fpr, tpr, thresholds, percentiles=[0.99], verbose = False)[0])\n",
    "\n",
    "        # # # # # # # # # # # # # # # # # # # # # # # # #\n",
    "        # 2. Reconstruction Loss AUC & elimination rates\n",
    "        # # # # # # # # # # # # # # # # # # # # # # # # #\n",
    "\n",
    "        y_anom = motion_losses[motion_labels == 1]\n",
    "        y_norm = motion_losses[motion_labels == 0]\n",
    "        # likelihoods do not need to be inverted (anomalies have higher reconstruction loss)\n",
    "        X = np.concatenate([y_norm, y_anom]).reshape((-1, 1))\n",
    "        y = np.concatenate([-np.ones_like(y_norm), np.ones_like(y_anom)])\n",
    "\n",
    "        fpr, tpr, thresholds = roc_curve(y, X[:,0])\n",
    "        loss_scores[\"auc\"].append(auc(fpr, tpr))\n",
    "        loss_scores[\"tnr90\"].append(get_percentiles(fpr, tpr, thresholds, percentiles=[0.9], verbose = False)[0])\n",
    "        loss_scores[\"tnr95\"].append(get_percentiles(fpr, tpr, thresholds, percentiles=[0.95], verbose = False)[0])\n",
    "        loss_scores[\"tnr99\"].append(get_percentiles(fpr, tpr, thresholds, percentiles=[0.99], verbose = False)[0])\n",
    "    return kde_scores, loss_scores"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plotting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_boxes(values, color=\"darkgoldenrod\", label_below=[], colors=None):\n",
    "    bp_dict = plt.boxplot(values, medianprops={\"linewidth\": 1.5, \"color\": color}, widths=0.68)\n",
    "\n",
    "    # Add labels to medians\n",
    "    for i, line in enumerate(bp_dict['medians']):\n",
    "        # get position data for median line\n",
    "        x, y = line.get_xydata()[1] # top of median line\n",
    "        # custom color\n",
    "        if colors is not None:\n",
    "            color = colors[i]\n",
    "        line.set_color(color)\n",
    "        # overlay median value\n",
    "        if i in label_below:\n",
    "            plt.text(x - 0.34, y - 0.0006, f\"{y:.4f}\"[1:], verticalalignment=\"top\", horizontalalignment=\"center\", color=color) # draw below, centered\n",
    "        else:\n",
    "            plt.text(x - 0.34, y + 0.0005, f\"{y:.4f}\"[1:], verticalalignment=\"bottom\", horizontalalignment=\"center\", color=color) # draw above, centered\n",
    "\n",
    "def plot_scores(configs, scores, metric=\"auc\", savefile=None, figsize=(15, 10), label_below=[], xlabel=None, colors=None):\n",
    "    plt.figure(figsize=figsize)\n",
    "\n",
    "    plot_boxes([kde_scores[metric] for kde_scores, loss_scores in scores], label_below=label_below, colors=colors)\n",
    "    # plot_boxes([loss_scores[metric] for kde_scores, loss_scores in scores], color=\"blue\")\n",
    "\n",
    "    # Add labels\n",
    "    plt.xticks(np.arange(1, len(configs) + 1), [config[\"label\"] for config in configs])\n",
    "    if xlabel is not None:\n",
    "        plt.xlabel(xlabel)\n",
    "    plt.grid(True, ls=\"dotted\", lw=0.5)\n",
    "    # plt.ylim((0.7, 0.9))\n",
    "    # plt.legend()\n",
    "    if savefile is not None:\n",
    "        plt.savefig(savefile, bbox_inches=\"tight\")\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Plot result\n",
    "## By latent features\n",
    "Choose a metric (\"auc\", \"tnr90\", \"tnr95\", \"tnr99\") and optionally an output file for plotting."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Session 'Beaver_01' at folder: /home/kleinsteuber/vscode/ResizedSessions256_NoBackup/VIELAAS_Spring_Session01-VIELAAS_Beaver_01\n",
      "Loaded scans.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10/10 [00:00<00:00, 21.04it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Session 'Beaver_01' at folder: /home/kleinsteuber/vscode/ResizedSessions256_NoBackup/VIELAAS_Spring_Session01-VIELAAS_Beaver_01\n",
      "Loaded scans.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10/10 [00:00<00:00, 15.49it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Session 'Beaver_01' at folder: /home/kleinsteuber/vscode/ResizedSessions256_NoBackup/VIELAAS_Spring_Session01-VIELAAS_Beaver_01\n",
      "Loaded scans.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10/10 [00:00<00:00, 14.54it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Session 'Beaver_01' at folder: /home/kleinsteuber/vscode/ResizedSessions256_NoBackup/VIELAAS_Spring_Session01-VIELAAS_Beaver_01\n",
      "Loaded scans.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10/10 [00:01<00:00,  8.92it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Session 'Beaver_01' at folder: /home/kleinsteuber/vscode/ResizedSessions256_NoBackup/VIELAAS_Spring_Session01-VIELAAS_Beaver_01\n",
      "Loaded scans.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10/10 [00:01<00:00,  5.40it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Session 'Beaver_01' at folder: /home/kleinsteuber/vscode/ResizedSessions256_NoBackup/VIELAAS_Spring_Session01-VIELAAS_Beaver_01\n",
      "Loaded scans.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10/10 [00:03<00:00,  2.69it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Session 'Beaver_01' at folder: /home/kleinsteuber/vscode/ResizedSessions256_NoBackup/VIELAAS_Spring_Session01-VIELAAS_Beaver_01\n",
      "Loaded scans.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10/10 [00:02<00:00,  3.56it/s]\n"
     ]
    }
   ],
   "source": [
    "scores = [get_scores(**config) for config in CONFIGS_BY_LATENT_FEATURES]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_scores(CONFIGS_BY_LATENT_FEATURES, scores, metric=\"auc\", savefile=\"plots/approach4/boxplot_kde_latentfeatures.pdf\", figsize=(6, 8), label_below=[0], xlabel=\"Latent features\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAHgCAYAAAC1lJYGAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAA98klEQVR4nO3deZxcZZno8d9Tqa7q7uqE7AESMCxZCSASgkjUoAiRRcBtBES9k1xEhfHqOFz94PUyOtxBXMaZUYyjuM0oMIoiQlhEAUXZEiAhIUASlhACSToJnXR1V3VX13v/qOq2k3R6qber3/Ocer6fT3/SS3XV77w5eVJ9uuqUOOcwxhijXyJ0gDHGmOFhA90YY2LCBroxxsSEDXRjjIkJG+jGGBMTNtCNMSYmkqFueOLEiW769OlVu/6uri5GjRpVteuvNusPS3O/5naw/oGsXLmy2Tk3qa+vBRvo06dPZ8WKFVW7/mw2SyaTqdr1V5v1h6W5X3M7WP9AROSlA30ttodcEgndm2b9YWnu19wO1u9128Fuucqy2WzoBC/WH5bmfs3tYP0+YjvQjTGm1sR2oKfT6dAJXqw/LM39mtvB+n3EdqAnk8F+3zssrD8szf2a28H6fcR2oNtxuLCsPxzN7WD9PmI70I0xptbEdqCnUqnQCV6sPyzN/Zrbwfp92ECPKOsPS3O/5nawfh+xHeitra2hE7xYf1ia+zW3g/X7iO1AN8aYWhPbgW4/toVl/eFobgfr9xHbgW5PTgjL+sPR3A7W7yO2A92Ow4Vl/eFobgfr9xHbge6cC53gxfrD0tyvuR2s30dsB3pdXV3oBC/WH5bmfs3tYP0+YjvQ7ThcWNYfjuZ2sH4fsR3obW1toRO8WH9Ymvs1t4P1+9B9WrN+FIvF0AlerD+sKPeLyLBcT1SPVUd57QcjZH9sB7qdgjMs66+egQaxiER2WA9GlNd+MOz0uVVQX18fOsGL9YelvV8z7Wsfsj+2A92Ow4Vl/aZS2tc+ZH9sB7odhwvL+k2ltK99yP7YDvREQvemWX9Y2vs10772Ift1r1w/GhsbQyd4sf6wtPdrpn3tQ/bHdqDncrnQCV6sPyzt/ZppX/uQ/bEd6IVCIXSCF+sPS3u/ZtrXPmR/bAe6HYcLy/pNpbSvvR1DrwI7DheW9ZtKaV97O4ZeBfl8PnSCF+sPS3u/ZtrXPmR/bAd6Z2dn6AQv1h+W9n7NtK99yP7YDvThOoFRKNYflvZ+zbSvfcj+2A70pqam0AlerD8s7f2aaV/7kP2xHeh2HC4s6zeV0r72dgy9Cjo6OkIneLH+sLT3a6Z97UP2x3agG2NMrYntQLfjcGFZv6mU9rW3Y+hVYD+2hWX9plLa194OuVSB7RRhWb+plPa1t4FujDHGW2wHeiaTCZ3gxfrD0t6vmfa1D9kf24Fup+AMy/pNpbSvvZ0+twrsyQlhWb+plPa1tycWGWOM8RbbgW7H4cKyflMp7Wtvx9CroFgshk7wYv1hae/XTPvah+yP7UBvb28PneDF+sPS3q+Z9rUP2R/bgW6MMbUmtgPdXpcwLOs3ldK+9vaaosYYY7zFdqC3tbWFTvBi/WFp79dM+9qH7I/tQDfGmFoT24He0NAQOsGL9YelvV8z7Wsfsj+2Az2R0L1p1h+W9n7NtK99yH7dK9ePbDYbOsGL9YelvV8z7Wsfsj+2A90YY2pNbAd6Op0OneDF+sPS3q+Z9rUP2R/bgZ5MJkMneLH+sLT3a6Z97UP2x3ag23G4sKzfVEr72tsxdGOMMd5iO9BTqVToBC/WH5b2fs20r33IfhvoEWX9YWnv10z72ttAr4LW1tbQCV6sPyzt/ZppX/uQ/bEd6MYYU2tiO9Dtx7awrN9USvva2yGXKrAnJ4Rl/aZS2tfenlhUBXYcLizrN5XSvvaRP4YuIotF5FkR2SAinz/AZRaJyJMislZEHhjezKFzzoVO8GL9YWnv10z72ofsH/A5qiIyCvgO8C5gM/CYiNzmnHu612XGAtcDi51zm0RkcpV6B62uri50ghfrD0t7v2ba1z5k/2DuoS8ANjjnnnfOdQA3Aeftc5mLgF855zYBOOe2DW/m0NlxuLCs31RK+9pH/Rj6VODlXh9vLn+ut5nAOBG5X0RWishHhiuwUva6hGFZv6mU9rUP2T+Y04JJH5/b9yBREjgReCfQADwkIg87557b64pELgUuBZg2bRq5XI6Ojg46OjoAyGQyFAoF8vl8z8fFYpH29nYAGhsbgb8uWENDA4lEoudkOOl0mmQySTabpaWlhWQySSqV6vklRSqVIp1O09rainOOuro60uk0bW1tFItFkskk9fX1PR8nEgkaGxvJ5XIUCoWej/P5PJ2dnYgITU1N5PP5nm1oamoalm3atWsXxWJxr23q3gYN29R9ewP9PUV1m/L5PNlstqJ9L/Q2ATQ3Nw/rv6eR3KY9e/b07DvD9e9pJLcpl8vR0tJStRnRHxnoAL6InAJc7Zw7s/zxFwCcc//c6zKfB+qdc1eXP74BuMs594sDXe/8+fPdihUrBgys1Ouvv87YsWOrdv3VZv1hae4XEdW/WNS89lD9fhFZ6Zyb39fXBnPI5TFghogcISIp4EPAbftc5jfAW0UkKSKNwMnAOp9oX/X19SFv3pv1h6W9XzPtax+yf8CB7pwrAJcDd1Ma0v/tnFsrIpeJyGXly6wD7gJWA48CP3DOrale9sDsOFxY1m8qpX3to34MHefccmD5Pp9bts/HXwO+NnxpfrqPwWll/WFp79dM+9qH7Nf9Wk/9SCR0Pwk26v0iff2ufGiifJw36uuvmfZ9J8r9sR3o3b8Z1irq/YP4ZXqkB/ZAor7+mmnfd6LcH9u7IblcLnSCF+392tn6G41iO9ALhULoBC/a+7Wz9TcaxXagaz8Gqr1fO1t/o1Fs91rtx0C192tn6280iu1A734arVba+7Wz9TcaxXagd3Z2hk7wor1fO1t/o1FsB/pwPFY0JO392tn6G41iO9CbmppCJ3jR3q+drb/RKLYDXfsxUO392tn6G41iO9C7zzWslfZ+7Wz9jUaxHejGGFNrYjvQtR8D1d6vna2/0Si2A137j8za+7Wz9Tca2UCPKO392tn6G41iO9CNMabWxHagZzKZ0AleQvePHz8eEan4DfD6/vHjxwfd/tDrb0wlYvsCF9pPfxq6f9euXV4n6d+zZw+jR4+u+PtDP1Mz9PobU4nY3kPX/sQQ6w9Le7+pTbEd6MYYU2tiO9C1HwO1/rC095vaFNuBXiwWQyd4sf6wtPeb2hTbgd7e3h46wYv1h6W939Sm2A50Y4ypNbEd6NpfE9L6w9Leb2pTbAe6McbUmtgO9La2ttAJXqw/LO39pjbFdqAbY0ytie1Ab2hoCJ3gxfrDCtkf+jw6UTiXjqlMbM/lkkjo/r/K+sMK2e97Hp329nbv/5BCn0vHVEb3v7p+ZLPZ0AlerD8szf2a242f2A50Y4ypNbEd6Ol0OnSCF+sPS3O/5nbjJ7YDPZnU/esB6w9Lc7/mduMntgNd+3FE66+ewTzKo7GxcVCPJomiKK+9qS77r9zUnME8gkREvB5pYkwIsb2HnkqlQid4sX5TKVv72mUDPaKs31TK1r52xXagt7a2hk7wYv2mUrb2tSu2A90YY2pNbAe69h87rd9UKvTaaz8XTeh+n/bYPspF+5Mrotj/1O0fZ9emB0mmxwAw75zvMWbKcWx77nbW//EriCSQRJLZp3+Vxskn0lXI8eh/nUmxK48rFjh41vkc/bYvArB762qevuvTFAs5JJFkzpn/wthD54fcvNgIve/4nosmn897b4PPw0pD9/u0x3agt7a2Bt+xfUS1f+Y7/omDZ1+w1+fGT1/EW2acjYiwZ9saVv36Eua8/17Gjx/PSRfdQTLVRLGrk0f/811MPOoMxk5dwHN/+CJHLfwCk446g+0b7ua5+77IgovvCrRV8RLVfWewrL9ysR3o2h9DrKk/mWrqeb+rIwvlx3CLSM/XXLGTYrETKN/7EKGQ3w1AId9CuumQkc6OLU37Tl+sv3KxHeh1dXWhE7xEtX/9A19m44PXMmH6ImYu+jKJZOmeyNZnb2P9/f+XfFszJ37gl0i53xW7eOhHC2nb9TyHnXgpY6eeBMDs07/KypvP57k/XIVzRU7+yO+DbVPcRHXfGSzrr5yE+t9k/vz5bsWKFVW7/lwuR319fdWuv9pC9/f1TMl862ukMlNwXR2svfMKGsYdwdELv7DXZXZuepCNf76WYy/45V79nbnXeeKWC5lzxtcZPekY1t3zOcYdvpCDZ5/Pa+tu4eUnf8RJF97e7+2PpJC373vbw7Hv+DRYv1//QLcvIiudc33+wim2j3LR/pqQUexPNx2MiJBIppl63IfZvWXlfpcZf/hC2ne9QMuOzXt9vq5+LOMPfyvNz98LwJY1P2fKrPMAmDL7vbT0cV2mMlHcd4bC+isX20MuxWIxdIKXKPbnW18j3XQwzjm2Pnc7TZPmApDduZHGcUciIux+7UmKXR0kUgfR0bYdSdRRVz+Wrs52drx4H0e8+bNA6T+HXZv+xPg3vI2dL91PZvxRITct8g70CKMta27mhYe/CZR+lzHnzG9RHHUI2R3PserWj/Z8f9vrL3L0W7/I9AWfoqN9J6tv/SjtLZtoOOhwjj//p9Q1jAuyXX2J4r4/FCH7YzvQtZ9CNEr9K29+L8ec9R2e+u0SOtqawTlGTzmOGYv/FYCtz/6GLWt+TiJRRyLZwPHn/wSpqyPfupmnbr8UV+wCV2TKnPcyeca7ATjm3d/mmXuvpFgsMGpUPXMX/3vITVShr0cYNYx9Awsuvou6hnFs33gPT995BbPPv5XM2Jm8ZclDQOn3GPd/ewZTZp0LwAsPfZPx0xdx5Cl/z/MPfYPnH/4ms077yohvz4FEad+vRMh+3SvXD83HzyFa/Sf+za8AOOmi5X1+/chTPsuRp3x2r8/lcjnqx87jLX/7lz6/Z9xhb+GU//Hg8IbWoHHT3tzz/thDTyK355X99p0dL95P49gjaTjocAC2rb+DBRffCcDUYy/m0Z+9O1IDPUr7fiVC9sd2oLe1taneMUL2r/vdlVy3tJ5Hf7a44usodBZI1lW+e338bHum6b4O9AijbptX/5SJR52x377z2rpfcvDc9/d83JHdRrrpYKB06KujbfvIbMAg2b/dysX2l6J2HC4sh+7HEkfNzEX/yMJLH+eUj/2RzvZdPF8+bt5tx0sP8MqqnzBz0Zf32neKXR1sW38HB8+5YN+rjCzt+74dQ6+CREL3/1Uh++e86zrmnvE13Pcrf+bmzp07vc5JcfKHhWUVf3f8dN+jlvIjjF585N96vrZn2xrWLr+cEz/4K1KNE0jkdvZ8rXnjPYyZ8kbSmSk9n0tlJvf8gjvf+hqpxkkjtyGDYP92PW472C1XWWNjY+gEL9Zvesu3vgaw3yOM2lte5olbLuLYc79PZsIMYO+1f/XpX3DIMR/Y67omzziLV576GQCvPPUzJs84eyQ2YdC07zsh+2N7Dz30E3N8Wb+BgR9htPHP19KZ28m6uz8DUDrR2QW/pb6+nq7ONna8cB9zF//bXtd5xJs/y6pbP8Irq35K/ZhpHH/Bf474dvVH+74Tsj+2zxRtbm5m4sSJVbv+agvd7/tsOd9+e6ZouLX3bbD+6u77/T1TNLb30O04XFja+0MZrkcYPe/xCCMI+ygj7fuOHUOvAjsOF5b2fs0So3T/s9a+74Tsj+0hl5aWFg466KCqXX+1he73/bHTt98OuYRbe98G66/uvl+TJ+fq7OwMneDF+k2ltK+99VcutgPd52WcosD6TaW0r731Vy62A72pqWngC0WY9ZtKaV97669cbAd6Pp8PneDF+k2ltK+99VcutgO9o6MjdIIX6zeV0r721l+52D4O3RgTHQd6gQ7nHM/87h/YvvEeRtU1cOw534PkNKD0soVrl3+K1u1PgwjzzvouY6edzO6tq3n6rk9TLORKz4w9818Ye2ifD/qovX7nXJC3E0880VVTe3t7Va+/2kL3l3aNyvn2+96+r5C3H3rtfRv6+t7Vv73UvbruV/t9ftv6u9yKm853xWLR7dr8iHvoR2/v6V992/90Lz/xI+ecc12FvOto3+Wcc+6xn5/rtm24u+f7H/mvM2PVP1A7sMIdYK7aIZeIsn5TKU1rv2397Rw670JEhLFTF9CZb6F118sU8rvZ9fKfmXp86WX0EqNS1NWPLX2TCIX8bgAK+RbSTYcEqo9ef2wPuWjaqfti/aZSUV37vl6gI7/nVerHTOu5TP3oQ8m+/jKpRDt1jRNZc8dl7Nn2FGMOPoHZp19HMpVh9ulfZeXN5/PcH67CuSInf+T31l8W23voxpjoONALdPT9QiiCKxbY89qTHHbCUt7yt39hVF0jLzz0DQBefvwHzHrntbz98meZffq1rFn+Sesvi+1Az2QyoRO8WL+pVBTXPt10MCJCovwCHbu3rARK92hzuzf3XC63ZwsHTZpO/eippMdMZezUkwA4ePb57N66CoAta37OlFnnATBl9ntpKV+X9cd4oBcKhdAJXqzfVCqKa3+gF+iYPONstqy5Eeccr7/yKMn0GEalJ5JumkL96KlkdzwHlF7oumnibKA0XHdt+hMAO1+6n8z4o6y/LLbH0PP5PKNHjw6dUTHrN5WK0toP9AIdE486k+0b7+ZPy45jVF0D885e1tM/54xvsPq2JRS7OmgcewTzzv4uAMe8+9s8c++VFIsFRo2qZ+7if7f+stiebTH0C0T4Ct0f9ZP8V5vmsy3W+gtE+DaE7rcXuOhDFI8jDoX1m0qFXPvheIGOYleR5z3P6V7pC3REod/nxUViewy9WCyGTvBi/aZS2tc+3M9lwyNkvx1yiajQ/VH/sbPa7JCL3kMWvg2h++0FLowxxgxuoIvIYhF5VkQ2iMjn+/j6IhFpEZEny29fGv7UobHXJfQnIhW/TZo0yev7x40bF3rz1YrCvuPD+is34C9FRWQU8B3gXcBm4DERuc059/Q+F/2Tc+6cKjSaAHwPN4Q+ZGJMLRrMPfQFwAbn3PPOuQ7gJuC86mb5a2trC53gRXu/CUf7vmP9lRvMwxanAi/3+ngzcHIflztFRFYBW4DPOefW7nsBEbkUuBRg2rRp5HI5Ojo6ek4mlMlkKBQKPa/4kclkKBaLtLe3A3/9UaZ7wRoaGkgkEmSzWQDS6TTJZJJsNktLSwupVIpUKkVraysAqVSKdDpNa2srzjnq6upIp9O0tbVRLBZJJpPU19f3fJxIJGhsbCSXy1EoFHo+zufzdHZ2IiI0NTWRz+d7tqGpqWlYtqmlpWW/bereBi3b1NzcPODfU1S3CSCbzVa07/luE8DOnTsr3qbufcf339OePXsq2qbuv/tK/566b89n3+tuqOTvCWDHjh0V73vdfwfV2vf6M+CjXETkA8CZzrml5Y8vARY4567odZkxQNE51yoiZwH/6pyb0d/1VvtRLtlsVvVjobX3az/kovlRLsOx74R8lEit91f7US6bgcN6fTyN0r3wHs653c651vL7y4E6EQn6mMFEQvcDeLT3m3C07zvW73Hbg7jMY8AMETlCRFLAh4Dbel9ARA4WESm/v6B8vTuGO3Yoev/YppH2fhOO9n3H+is34DF051xBRC4H7gZGAT90zq0VkcvKX18GvB/4hIgUgHbgQ07zz9vGGKPQoM7lUj6Msnyfzy3r9f63gW8Pb5qf7l/OaKW934Sjfd+x/srpPljVj2RS93nHtPebcLTvO9ZfudgOdDsOZ2qV9n3H+isX24FujDG1JrYDvfsJAlpp7zfhaN93rL9yNtAjSnu/CUf7vmP9lYvtQO9+Kq9W2vtNONr3HeuvnO5fJxsTU+Xn6QVjpy/WKbYD3X5sM1rV+qmLte/7dsilCuzJCcbopH3ftycWVYEdh6td48eP93q1pO7DHT7fP378+MCroJf2fd+OoVeB5h85QX9/SLt27fJev+F4oV9TGe37fsj+2N5Dr6urC53gRXu/drb+4Whf+5D9sR3odhzO+LD1D0f72tsx9Cqw1yU0Pmz9w9G+9iH7YzvQi8Vi6AQv2vu1s/UPR/vah+yP7UC3U3AaH7b+4Whfezt9bhXU19eHTvCivV87W/9wtK99yH7d/xX2o62tTfWOob1fO1t/P6Eftul76oKQ/T7tsR3odhzO+LD1r5z2Uxdo7o/tIZdEQvemae/XztbfaBTbvbaxsTF0ghft/drZ+huNYjvQc7lc6AQv2vu1s/U3GsV2oBcKhdAJXrT3a2frbzSK7UDXfgxUe792tv5Go9jutdqPgWrv187W32gU24Gez+dDJ3jR3q+drb/RKLYDvbOzM3SCF+392tn6G41iO9BDP1PNl/Z+7Wz9jUaxHehNTU2hE7xo79fO1t9oFNuBrv0YaNT7h+M1OaMs6utvTF9iO9A7OjpCJ3iJer9zrt+37du3D3iZKIv6+hvTl9ienMuYvqy75+95ZfV/cfrntu71+ZYtK3n4p6dx/Pk/4eDZFwDwwPVzSaaaEBmFJJKc8j/+BMDurU/x9F2fpquzlYaD3sBx77mBZHrMiG+LMfuK7UDXfgzU+odfy6uP05lr2e/zrtjFc/f/HyYecXrP57r7T7poOanGiXtdfu3yTzHrndcw/vC3snnVT3nh4W8x4+1fqm68MYNgh1wiyvqHlyt28ewfrmLWO/5pv6+9tGIZU2adRyozqedz/fVnd65n3GELAZhwxDvY+uxvhj/YmArYQI8o6x9em1YuY/KMs0k3HbzX53N7trDtuds47ISle32+o6MDQVhx03k89KOFvPzED3u+NnrSXLavvwOArc/8mtyeV6q/AcYMQmwPuRjTLbfnVV575lZOuvjO/b72zL1XMvO0ryCJUft9bcEl91I/+hDy2W2suOk9ZCbMZPzhCznm7Ot55p5/YOOfr2XS0WeRSKRGYjOMGVBsB3omkwmd4MX6h8+erato27WRPy07DoCuzjb++N3jeNsnVrP71SdY9ZuPAdDZtoPmjXcjiSRjDjud+oYGANKZyUyZeS4tr65k/OELaZowi/kX3gZAdsd6tm+8O8h2GbOv2A507ac/tf7hM+noxZz2d8/3fHzv16fwtk+sBuBtn1zb8/mnbv84k45ezJSZ5/L6zq3UJQok06MpdGTZ8cIfOPLUzwOQz24jnZmMc0We/8t1HHbCkpHdIGMOILYDPZ/PM3r06NAZFbP+4bHy5vdyzFnfoX70IUP6vuzrr7D2158EwBULHDL3g0w66l0AvPb0L9i08vsATJn1HqYed8nwRhtTIQn1BI/58+e7FStWVO36m5ubmThx4sAXjCjrr9xwvEivb3/IFwoO/SLLvqx/wOtf6Zyb39fXYnsPPUrHcCth/ZVZ97sruW5pPY/+bLHX9RS7ijw/qvIHgX38bPtFqRl5sX3YYrFYDJ3gxfrD0nv/0NQyO+QSUdZfOTvkYocsQgp5yCW299CNMabWxHaga39NSOsPS3u/qU2xHejGGFNrYjvQ29raQid4sf6wtPeb2hTbgW6MMbUmtgO9oXweDq2sPyzt/aY2xXagJxK6N836w9Leb2pTbPfabDYbOsGL9Yelvd/UptgOdGOMqTWxPZdLOp0OneDF+v2ISNDbHzduXNDbN7UptgM9mdS9adZfueF42rX2p5+b2hTbQy7aj4FavzFmqGI70I0xptbEdqCnUrrPR239xpihsoEeUdZvjBmq2A701tbW0AlerN8YM1SxHejGGFNrYjvQtf/Ib/3GmKGK7UAP/cQWX9ZvjBmq2A507cdwrd8YM1SxHejan+Vn/caYodL9/PJ+1NXVhU7o13CcayTKQzPq629MHMV2oEf9GO5Aw1j7uUSivv7GxFFsD7nYa0KGZetvzMiL7UAvFouhE2qarb8xIy+2A1376We1s/U3ZuTFdqDX19eHTqhptv7GjLzYDnQ7hhuWrb8xIy+2A92O4YZl62/MyIvtQE8kYrtpKtj6GzPyYvuvrrGxMXRCTbP1N2bkxXag53K50Ak1zdbfmJEX24FeKBRCJ9Q0W39jRl5sHywc+hju+PHj2bVrl9d1+JzvZdy4cezcudPr9n2EXv/+DHZdB7pcqFMzDKZ/MJfRfGoJ07dB/asTkcUi8qyIbBCRz/dzuZNEpEtE3j98iZUJfQx3165dOOcqfmtvb/f6ft//THyFXv/+DNf6R7V/sPuOiZ8BB7qIjAK+A7wbmAtcKCJzD3C5rwJ3D3dkJfL5fOgEL9YfluZ+ze3Gz2DuoS8ANjjnnnfOdQA3Aef1cbkrgFuAbcPYV7HOzs7QCV6sPyzN/ZrbjZ/BHEOfCrzc6+PNwMm9LyAiU4ELgHcAJx3oikTkUuBSgGnTppHL5ejo6KCjowOATCZDoVDouYeRyWQoFou0t7cDf/0xvvtZiA0NDSQSCbLZLFA6ZWsymSSbzdLS0kIqlSKVSvW8ek4qlSKdTtPa2opzjrq6OtLpNG1tbRSLRZLJJPX19T0fJxIJGhsbyeVyFAqFno/z+TydnZ2ICE1NTeTz+Z5taGpq6nm/ubm54m1qaWnZb5u6t2Ew29R9+8O5TUP5e8rn8zQ3Nw/49zSUbarG39OBtqmjo4NsNlvRvhd6m7r3neH89zSS29S971a674XeJoCWlpZh/ffUe5v6I4M4L/cHgDOdc0vLH18CLHDOXdHrMr8AvuGce1hEfgzc7pz7ZX/XO3/+fLdixYoBAyuVz+eDnpPb93zmvv2hz6ceev19ae7X3A7h911f1e4XkZXOufl9fW0w99A3A4f1+ngasGWfy8wHbir/Zn0icJaIFJxztw49d3ho36mtPyzN/ZrbjZ/BDPTHgBkicgTwCvAh4KLeF3DOHdH9fq976LcOX+bQdf84o5X1h6W5X3O78TPgQHfOFUTkckqPXhkF/NA5t1ZELit/fVmVG40xxgzCoJ5Y5JxbDizf53N9DnLn3Mf8s/x1/3JFK+sPS3O/5nbjJ7bPFO3o6Aj2Igvrfncl1y2t59GfLa74OroKXYxKjqr4+z9+dqri7x0OIdd/OGju19xu/MR6oGtWdEVGUflAD037+mvuj3r7cJy6IOSjYKLcH9uBHtKcd13H3DO+hvv+XXt9fs0dn6TltcfBOTLjj2beOd8jmWqis30Xa5Z/grZdL5BI1jPv7OvJyxQmTpzImjs+wfYNd5JqnMSp//OxnuvavfUpnr7r03R1ttJw0Bs47j03kEyP6fn6yR8W7JcbJooGGmbNzc1MnDhxhGqGLsr90T2DkqdMJhM6YT+zT7+WU5c8zKlLH6F+zGFsWvk9AJ5/6OuMnnwcpy59hGPP/Q+e+d2VPf2HHnsxJ/7Nrftd19rln2Lmaf/IqUsfZfLMc3nh4W+N4JYMLIrrPxSa+zW3g/X7iO1Aj+LpW7vvQTvnKBbagdKPZa3NzzBh+iIAmibMor1lE227XwVg/OELqasft991ZXeuZ9xhCwGYcMQ72Prsb6q/AUMQxfUfCs39mtvB+n3EdqBH9QRFT91+Gff/25G07niON8y/DIDRk49l67O3AfD6lhXkWjbRuvOlfq9n9KS5bF9/BwBbn/k1uT2vVDd8iKK6/oOluV9zO1i/j9gO9Kg69pxlLLpiA00TZvHaulsAOPKUz9KZe52/3HAKm1YsY/SU45FE/78QPebs69m08j946EcLKeT3kEiEfVSLMSa82P5SNMrH4SQxioPnvo8XH/5Xph53Ccn0GI49p/QrTOccf/zuMYybMrvf62iaMIv5F5bu1Wd3rGf7xkictbhHlNd/MDT3a24H6/cR23voxWIxdMJenHNkd27seX/7+jvJTJgJQGfudYpdpYeabV71Y8YfdiqJuv53inx2W/m6ijz/l+s47IQlVawfuqit/1Bp7tfcDtbvI7b30Nvb2yPzP/3Km9/LMWd9hzW3f5xCx25wjtGTj2Xu4m8BkG1+lqduvxSRBJmJs5l31vW0ZEv9q279GDs3/YnO9h3c/+2ZHP3Wq5h2/Ed57elfsGnl9wGYMus9TD3ukoBbuL8orX8lNPdrbgfr9zHg6XOrpdqnzw39WFbfU2j69oc+BWno9feluV9zO1j/QPo7fW5sD7lE+TUtB8P6w9Lcr7kdrN9HbAe6McbUGrUDXUT6fctkMgNeJsoG83JTUWb94WhuB+v3ofaXooN46TzVL2NljDFDpfYeetw1NDSETvBi/eFobgfr92EDPaK6Xz1cK+sPR3M7WL/XbQe7ZdOvbDYbOsGL9YejuR2s34cNdGOMiQkb6BGVTqdDJ3ix/nA0t4P1+7CBHlHJpNoHIAHWH5LmdrB+HzbQI8qOI4aluV9zO1i/DxvoxhgTEzbQIyqV0v2CFdYfjuZ2sH4fNtAjynbqsDT3a24H6/dhAz2iWltbQyd4sf5wNLeD9fuwgW6MMTFhAz2i7MfOsDT3a24H6/dhAz2i7MkVYWnu19wO1u/DBnpE2XHEsDT3a24H6/dhAz2itJ/L3frD0dwO1u/DBnpE1dXVhU7wYv3haG4H6/dhAz2i7DhiWJr7NbeD9fuwgR5R9rqKYWnu19wO1u/DBnpEFYvF0AlerD8cze1g/T5soEeUnUI0LM39mtvB+n3YQI+o+vr60AlerD8cze1g/T5soEeUHUcMS3O/5nawfh820CPKjiOGpblfcztYvw8b6BGVSOj+q7H+cDS3g/V73XawWzb9amxsDJ3gxfrD0dwO1u/DBnpE5XK50AlerD8cze1g/T5soEdUoVAIneDF+sPR3A7W78MGekTZccSwNPdrbgfr97rtYLds+mXHEcPS3K+5Hazfhw30iMrn86ETKnLjjTcyb948MpkM8+bN48YbbwydVBGt6w+628H6feh+jm2MdXZ2hk4YshtvvJGrrrqKG264gdmzZ/PMM8+wZMkSAC688MLAdUOjcf27aW4H6/dh99AjSkRCJwzZNddcww033MBpp51GKpXitNNO44YbbuCaa64JnTZkGte/m+Z2sH4fdg+9ikL+xY4bN27Eb3PdunUsXLgQgKamJgAWLlzIunXrRrzFV3e/Rprbwfp92D30KnHOeb35XsfOnTtHfJvnzJnDgw8+CPz1OOKDDz7InDlzRrzFl+bjuJrbwfp92EA3w+aqq65iyZIl3HfffWSzWe677z6WLFnCVVddFTptyDo6OkInVExzO1i/j8gechk/fjy7du3yug6fQx7jxo0Lci9Xs+5ffF5xxRWsW7eOOXPmcM0116j7hagxWkmoV6ieP3++W7FixQG/LiJer56dy+W8zkvse/u+Qt++L9/1D01zv+Z2sP6BiMhK59z8vr4W20Mu2n9s0077+mvu19wO1u/DBrqpCu3rr7lfcztYv4/YDnRjjKk1sR3omUwmdEJN077+mvs1t4P1+4jtQNd+Ck7ttK+/5n7N7WD9PiL5sMV1v7uS65bW8+jPFld8HZ2dndTV1VX8/R8/O1Xx95rSkytGjx4dOqNimvs1t4P1+4jtPXRjjKk1qh6HvuaOT9Ly2uPgHJnxRzPvnO+RTDWxZc3NvPDwNwFIppqYc+a3qBtzNA0NDQC4YhcP/fit1Dcdyps++EsA1j/wZbatvwORBKnGScw753vUjz6k39sfSaFv31d7e3vP+mukuV9zO1j/QGLzOPTZp1/LqUse5tSlj1A/5jA2rfweAA1j38CCi+/i1KWPcOSp/5un77yCYrHY830vrbiezIRZe13XEW/+X5y69BHesuQhJh29mI1//ucR3Za4673+Gmnu19wO1u9D1UBPpscApZNWFQvtQOmp/eOmvZm6htLZBcceehK5Pa/Q3t4OQG73K2zfcBfTjv9on9cF0NXZhqD7lJ1R073+Wmnu19wO1u8jkr8U7c9Tt19G88a7yUyczax37n+vevPqnzLxqDN6Pn7m3iuZedo/0dWxZ7/Lrn/garY8dSPJ9BhOunh5VbuNMabaVN1DBzj2nGUsumIDTRNm8dq6W/b62o6XHuCVVT9h5qIv09jYyLb1d5JqnMRBh5zQ53XNePvVvP3yZznkmL9h04rvjUR+zbDXhQxHcztYvw91Ax1AEqM4eO772PrMb3o+t2fbGtYuv5wT3nczqcYJALz+ysNs27CcB66fy6rffIwdLz3A6tuW7Hd9hxzzQbY++5v9Pm+MMZqoGejOObI7N/a8v339nWQmzASgveVlnrjlIo499/tkJswAoK2tjZmL/pFFlz/H2z/5NMef92MmvOHtHPeeGwDI7tzQc93b1t/Rc11meLS1tYVO8KK5X3M7WL8PFcfQV978Xo456zusuf3jFDp2g3OMnnwscxd/C4CNf76WztxO1t39GQAkkWTGub/u9zqfu/9LtO1YD5Kg4aDDmbv4X6u9GcYYU1WqHoc+FNls1uucCqEfBx769n35rn9omvs1t4P1DyQ2j0MfikQitpumgvb119yvuR2s3+u2g91ylWWz2dAJNU37+mvu19wO1u8jtgPdGGNqTWwHejqdDp1Q07Svv+Z+ze1g/T5iO9CTSRUP4Ikt7euvuV9zO1i/j9gOdO3H4bTTvv6a+zW3g/X7iO1AN8aYWhPbgZ5K2SsOhaR9/TX3a24H6/cxqIEuIotF5FkR2SAin+/j6+eJyGoReVJEVojIwuFPHRrtO4V22tdfc7/mdrB+HwMOdBEZBXwHeDcwF7hQRObuc7HfA8c7594I/C3wg2HuHLLW1tbQCTVN+/pr7tfcDtbvYzD30BcAG5xzzzvnOoCbgPN6X8A51+r++jz1DKD3OevGGKPUYB5fMxV4udfHm4GT972QiFwA/DMwGTi7rysSkUuBSwGmTZtGLpejo6ODjo4OADKZDIVCgXw+D5Re+aNYLPa8Akj3eYa7z2bW0NBAIpHo+a1yOp0mmUySzWbJZrOkUilSqVTP/5ipVIp0Ok1rayvOOerq6kin07S1tVEsFkkmk9TX1/dc/86dO2lsbCSXy1EoFEgkEjQ2NpLP5+ns7EREaGpqIp/P92xDU1NTv9uUyWQGtU0Azc3Ne21T9zZUuk3FYrFnG6q9TV1dXTQ3Nw/49xTVbSoWi2Sz2Yr2vdDb1N1Q6b4Xeps6Ojp69p3h+vc0ktsE0NLSUrUZ0Z8BT84lIh8AznTOLS1/fAmwwDl3xQEu/zbgS8650/u73mqfnCufz3s9wD/0ybFC374v3/UPTXO/5naw/oH4npxrM3BYr4+nAVsOdGHn3B+Bo0Rk4pAqh5n243DaaV9/zf2a28H6fQxmoD8GzBCRI0QkBXwIuK33BUTkaBGR8vtvAlLAjuGOHYqo37sVkX7fBnuZqIr6+g9Ec7/mdrB+HwMeQ3fOFUTkcuBuYBTwQ+fcWhG5rPz1ZcD7gI+ISCfQDvyNC/y3UldXF/LmBzTQ8rS0tHDQQQeNUM3wi/r6D0Rzv+Z2sH4fgzrpgHNuObB8n88t6/X+V4GvDm+aH83H4MD6Q9Pcr7kdrN9HbJ8paq9LGJb1h6O5HazfR2wHerFYDJ3gxfrD0tyvuR2s30dsB7qdgjMs6w9HcztYv4/YDvT6+vrQCV6sPyzN/Zrbwfp9xHag23G4sKw/HM3tYP0+YjvQ7ThcWNYfjuZ2sH4fsR3o3edU0Mr6w9Lcr7kdrN/rtoPdcpV1n9BGK+sPS3O/5nawfh+xHei5XC50ghfrD0tzv+Z2sH4fsR3ohUIhdIIX6w9Lc7/mdrB+H7Ed6HYcLizrD0dzO1i/120Hu+Uqs+NwYVl/OJrbwfp9xHagd7/6h1bWH5bmfs3tYP0+YjvQOzs7Qyd4sf6wNPdrbgfr9xHbgR71F4AYiPWHpblfcztYv4/YDvSmpqbQCV6sPyzN/Zrbwfp9xHag23G4sKw/HM3tYP0+YjvQOzo6Qid4sf6wNPdrbgfr9xHbgW6MMbUmtgPdjsOFZf3haG4H6/cR24FuP7aFZf3haG4H6/dhAz2irD8szf2a28H6fcR2oBtjTK2J7UDPZDKhE7xYf1ia+zW3g/X7iO1At1NwhmX94WhuB+v3EduBbk9OCMv6w9HcDtbvI7YD3Rhjak1sB7odhwvL+sPR3A7W7yO2A71YLIZO8GL9YWnu19wO1u8jtgO9vb09dIIX6w9Lc7/mdrB+H7Ed6MYYU2tiO9DtdQnDsv5wNLeD9fuI7UA3xphaE9uB3tbWFjrBi/WHpblfcztYv4/YDnRjjKk1sR3oDQ0NoRO8WH9Ymvs1t4P1+4jtQE8kdG+a9YeluV9zO1i/120Hu+VBEJGK3xobG72+f9y4cUG3PZvNBr19X9YfjuZ2sH4fyWC3PADnnNf3i4j3dRhjjCaRvodey9LpdOgEL9YfjuZ2sH4fNtAjKpmM7A9Pg2L94WhuB+v3YQM9ouw4Ylia+zW3g/X7sIFujDExYQM9olKpVOgEL9YfjuZ2sH4fNtAjynbqsDT3a24H6/dhAz2iWltbQyd4sf5wNLeD9fuwgW6MMTFhAz2i7MfOsDT3a24H6/dhAz2i7MkVYWnu19wO1u/DBnpE2XHEsDT3a24H6/dhAz2itJ+HxvrD0dwO1u/DBnpE1dXVhU7wYv3haG4H6/dhAz2i7DhiWJr7NbeD9fuwgR5R9rqKYWnu19wO1u/DBnpEFYvF0AlerD8cze1g/T5soEeUnUI0LM39mtvB+n3YQI+o+vr60AlerD8cze1g/T5soEeUHUcMS3O/5nawfh820CPKjiOGpblfcztYvw8b6BGVSOj+q7H+cDS3g/V73XawWzb9amxsDJ3gxfrD0dwO1u/DBnpE5XK50AlerD8cze1g/T5soEdUoVAIneDF+sPR3A7W78MGekTZccSwNPdrbgfr97rtYLds+mXHEcPS3K+5Hazfhw30iMrn86ETvFh/OJrbwfp92ECPqM7OztAJXqw/HM3tYP0+bKBHlIiETvBi/eFobgfr92EDPaKamppCJ3ix/nA0t4P1+7CBHlF2HDEszf2a28H6fdhAj6iOjo7QCV6sPxzN7WD9PmygG2NMTNhAjyg7jhiW5n7N7WD9PmygR5T92BmW5n7N7WD9PmygR5Tt1GFp7tfcDtbvwwa6McbEhA30iMpkMqETvFh/OJrbwfp9DGqgi8hiEXlWRDaIyOf7+PrFIrK6/PYXETl++FNri51CNCzN/Zrbwfp9DDjQRWQU8B3g3cBc4EIRmbvPxV4A3u6cOw74CvAfwx3aR1e/b4O9TFTZkyvC0tyvuR2s38dg7qEvADY45553znUANwHn9b6Ac+4vzrld5Q8fBqYNb+b+nHP9vm3fvn3AyxhjTJwkB3GZqcDLvT7eDJzcz+WXAHf29QURuRS4FGDatGnkcjk6Ojp6fiucyWQoFAo9/8NlMhmKxSLt7e3AX88z3NbWBkBDQwOJRIJsNgtAOp0mmUySzWbJ5/Ps3r2bVCpFa2srAKlUinQ6TWtrK8456urqSKfTtLW1USwWSSaT1NfX93ycSCRobGwkl8tRKBR6Ps7n83R2diIiNDU1kc/ne7ahqalpWLYpn8/T3Ny81zZ1b4OGbUokEjQ3Nw/49xTVbRo1ahTZbLaifS/0NnXvO8P572kkt6lYLPbsO8P172kktymZTNLS0lK1GdEfGeieqoh8ADjTObe0/PElwALn3BV9XPY04HpgoXNuR3/XO3/+fLdixYoBAyuVzWZV/3LF+sPS3K+5Hax/ICKy0jk3v6+vDeaQy2bgsF4fTwO29HEjxwE/AM4baJiPhO7/3bSy/rA092tuB+v3MZiB/hgwQ0SOEJEU8CHgtt4XEJHDgV8Blzjnnhv+TGOMMQMZ8Bi6c64gIpcDdwOjgB8659aKyGXlry8DvgRMAK4vP3qkcKAfCUaKvS5hWNYfjuZ2sH4fg/mlKM655cDyfT63rNf7S4Glw5tmjDFmKGL7TNHB/EY4yqw/LM39mtvB+n3EdqAbY0ytie1Ab2hoCJ3gxfrD0tyvuR2s30dsB3oioXvTrD8szf2a28H6vW472C1XWfczwbSy/rA092tuB+v3EduBbowxtSa2Az2dTodO8GL9YWnu19wO1u8jtgM9mRzUQ+wjy/rD0tyvuR2s30dsB7odhwvL+sPR3A7W7yO2A90YY2pNbAd6KpUKneDF+sPS3K+5Hazfhw30iLL+sDT3a24H6/cx4AtcVO2GRbYDL1XxJiYCzVW8/mqz/rA092tuB+sfyBucc5P6+kKwgV5tIrIi9Cl8fVh/WJr7NbeD9fuI7SEXY4ypNTbQjTEmJuI80P8jdIAn6w9Lc7/mdrD+isX2GLoxxtSaON9DN8aYmhKLgS4iPxSRbSKyZp/PXyEiz4rIWhG5LlRff0SkXkQeFZFV5c5/LH/+ayLyjIisFpFfi8jYwKkHJCJjReSX5d51InJKr699TkSciEwM2dhbX/vLgdZbROpE5Cci8lR5274QLLxMRA4TkfvKPWtF5NPlz18tIq+IyJPlt7N6fc9xIvJQ+fJPiUh9uC0AEXmx3PGkiKwof+4D5b6iiMzvddl3icjK8uVXisg7Rri1r/1lvIj8TkTWl/8cN9hWEblt31k1bJxz6t+AtwFvAtb0+txpwL1Auvzx5NCdB2gXoKn8fh3wCPBm4AwgWf78V4Gvhm7tZxt+Aiwtv58CxpbfPwy4m9LzDSaG7hxgf+lzvYGLgJvK7zcCLwLTA/cfAryp/P5o4DlgLnA18Lk+Lp8EVgPHlz+eAIwKvA0v7rtPAHOAWcD9wPxenz8BOLT8/jzglQjsL9cBny+///le+0u/rcB7gZ/3vq7hfIvFPXTn3B+Bnft8+hPAtc65fPky20Y8bBBcSWv5w7rym3PO3eOcK5Q//zAwLUjgAERkDKUd/gYA51yHc+718pf/BbgSiNQvavraX/pZbwdkRCQJNAAdwO6Rau2Lc+5V59zj5ff3AOuAqf18yxnAaufcqvL37HDOdVW/dGicc+ucc8/28fknnHNbyh+uBepFZMTOUXuA+XIepTsylP88v3zZA7aKSBPwWeCfqtUai4F+ADOBt4rIIyLygIicFDroQERklIg8CWwDfuece2Sfi/wtcOeIhw3OkcB24Eci8oSI/EBEMiLyHkr3TlYF7qtE7/X+JZAFXgU2AV93zu37jzsYEZlO6V5h9z5zefmw0Q+7DwNQ+rfgRORuEXlcRK4M0boPB9xTPixx6RC+733AE9131AKa4px7FUr/wQKT+7jMvq1fAb4BtFUrKs4DPQmMo3T44h+A/xYRCZvUN+dcl3PujZTuFS4QkXndXxORq4AC8LNAeQNJUvpx9LvOuRMoDb+rgauALwXsqkgf670A6AIOBY4A/l5EjgyUt5fyPb5bgP/lnNsNfBc4Cngjpf+AvlG+aBJYCFxc/vMCEXnniAfv7VTn3JuAdwOfEpG3DfQNInIMpcNhH692nK99W0XkjcDRzrlfV/N24zzQNwO/Kh/SeBQoUjrHQmSVD1XcDywGEJGPAucAF7vyAbgI2gxs7vVTxS8pDfgjgFUi8iKl/6geF5GDwyQOzgHW+yLgLudcZ/mw3Z+B4E9LF5E6SsP8Z865XwE457aW7xwUge9T+s8ISn9HDzjnmp1zbcBySn9HwXQfliiv6a/5a2ufRGRa+XIfcc5trH7hgLaKyCEA5T97DukeoPUU4MTyv4cHgZkicv9wR8V5oN8KvANARGZS+mVd5E74IyKTej2iogE4HXhGRBYD/xt4T/kfYSQ5514DXhaRWeVPvRN43Dk32Tk33Tk3ndJAeVP5spHUz3pvAt4hJRlKP/E9E6KxW/knzRuAdc65b/b6/CG9LnYB0P1IiruB40Sksfy7gLcDT49U777Kh+RGd79P6Rj/AR/1Uf73cQfwBefcn0ckcmC3AR8tv/9R4Ddw4Fbn3Hedc4eW/z0sBJ5zzi0a9qpq/KZ1pN+AGyn9iNlJaXgsoTTA/4vSjvI48I7QnQdoPw54gtKjENYAXyp/fgPwMvBk+W1Z6NZ+tuGNwIryNtwKjNvn6y8SrUe59LW/9LneQBPwC0q/4Hoa+IcI9C+kdAx6da/es4D/BJ4qf/424JBe3/Ph8jasAa4L3H8ksKr8tha4qvz5C8p/H3lgK3B3+fNfpHQo78lebyP2qLUD7C8TgN8D68t/jh9sKzCdKj3KxZ4paowxMRHnQy7GGFNTbKAbY0xM2EA3xpiYsIFujDExYQPdGGNiwga6CUpEWge+VM9lF4nIWzxua6yIfLKfr/9d+QyGQ35WrohMF5GLKm0zZjjYQDeaLAIqHujAWOCAA738tbOccxdXcN3TKT2rdEhEZFQFt2VMn2ygm8gRkXPLJ1V7QkTuFZEp5ZNQXQZ8pnwO7beWn2V7i4g8Vn47tfz9V5dPTnW/iDwvIn9XvuprgaPK3/+1fW5zGaUnvNwmIp8pP5vxh+XrfUJEzitfbrqI/Kl8kqvHe/3EcC2lk8E9Wf7+j4nIt3td/+0isqj8fquIfFlEHgFOEZEPS+mc+E+KyPfKJ2sbJSI/FpE1Ujq39meqtd4mRkI/683eavsNaO3jc+P468sjLgW+UX7/anqd75vSeaUXlt8/nNJT4bsv9xcgTen8PTsonZZ4Ov08Q49ez2gF/h/w4fL7YymdczxD6Zzo9eXPzwBWlN9fBNze67o+Bny718e3A4vK7zvgg+X35wC/BerKH18PfAQ4kdKZN7u/f2zovyt7i/5b0uP/AmOqZRpwc/ncJCnghQNc7nRgbq+TaI7pPkcIcIcrnbY0LyLbgClDbDgDeI+IfK78cT2l/zS2AN8unz2vi9KpaYeqi9KJtaB07psTgcfK29FA6URPvwWOFJF/p3RukHsquB1TY2ygmyj6d+Cbzrnbyocprj7A5RLAKc659t6fLA/G3ufL7mLo+7oA73P7vOCCiFxN6Twjx5dvP3eA7y+w9yHN3i/5lnN/fYEJAX7inNvvpe1E5HjgTOBTwAcpnafdmAOyY+gmig4CXim//9Fen99D6SXXut0DXN79Qflec3/2/f7+3A1c0X0OfRE5oVfbq650itpLgO5fau573S8CbxSRhIgcxoFPD/t74P0iMrl8O+NF5A1Seg3WhHPuFuD/EPh0t0YHG+gmtEYR2dzr7bOU7pH/QkT+xN6nPP4tpRdneFJE3gr8HTBfSq/Q8zSlX5oekHNuB/Dn8i8av9bfZSm9ukwdsFpKL+j7lfLnrwc+KiIPUzrcki1/fjVQkNKLfX+G0nnTX6B09sOvUzrjZ19NT1M6Q989IrIa+B2l1wydCtwvpVey+jEQ/MWpTfTZ2RaNMSYm7B66McbEhA10Y4yJCRvoxhgTEzbQjTEmJmygG2NMTNhAN8aYmLCBbowxMWED3RhjYuL/A1V13U3jWhuOAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 432x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_scores(CONFIGS_BY_LATENT_FEATURES, scores, metric=\"tnr95\", savefile=\"plots/approach4/boxplot_kde_latentfeatures_tnr95.pdf\", figsize=(6, 8), label_below=[0, 1], xlabel=\"Latent features\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Denoising autoencoders"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Session 'Beaver_01' at folder: /home/kleinsteuber/vscode/ResizedSessions256_NoBackup/VIELAAS_Spring_Session01-VIELAAS_Beaver_01\n",
      "Loaded scans.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10/10 [00:02<00:00,  3.48it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Session 'Beaver_01' at folder: /home/kleinsteuber/vscode/ResizedSessions256_NoBackup/VIELAAS_Spring_Session01-VIELAAS_Beaver_01\n",
      "Loaded scans.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10/10 [00:03<00:00,  2.83it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Session 'Beaver_01' at folder: /home/kleinsteuber/vscode/ResizedSessions256_NoBackup/VIELAAS_Spring_Session01-VIELAAS_Beaver_01\n",
      "Loaded scans.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10/10 [00:03<00:00,  2.74it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Session 'Beaver_01' at folder: /home/kleinsteuber/vscode/ResizedSessions256_NoBackup/VIELAAS_Spring_Session01-VIELAAS_Beaver_01\n",
      "Loaded scans.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10/10 [00:03<00:00,  2.99it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Session 'Beaver_01' at folder: /home/kleinsteuber/vscode/ResizedSessions256_NoBackup/VIELAAS_Spring_Session01-VIELAAS_Beaver_01\n",
      "Loaded scans.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10/10 [00:03<00:00,  2.64it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Session 'Beaver_01' at folder: /home/kleinsteuber/vscode/ResizedSessions256_NoBackup/VIELAAS_Spring_Session01-VIELAAS_Beaver_01\n",
      "Loaded scans.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10/10 [00:03<00:00,  2.90it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Session 'Beaver_01' at folder: /home/kleinsteuber/vscode/ResizedSessions256_NoBackup/VIELAAS_Spring_Session01-VIELAAS_Beaver_01\n",
      "Loaded scans.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10/10 [00:03<00:00,  2.58it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Session 'Beaver_01' at folder: /home/kleinsteuber/vscode/ResizedSessions256_NoBackup/VIELAAS_Spring_Session01-VIELAAS_Beaver_01\n",
      "Loaded scans.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10/10 [00:03<00:00,  2.65it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Session 'Beaver_01' at folder: /home/kleinsteuber/vscode/ResizedSessions256_NoBackup/VIELAAS_Spring_Session01-VIELAAS_Beaver_01\n",
      "Loaded scans.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10/10 [00:04<00:00,  2.11it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Session 'Beaver_01' at folder: /home/kleinsteuber/vscode/ResizedSessions256_NoBackup/VIELAAS_Spring_Session01-VIELAAS_Beaver_01\n",
      "Loaded scans.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10/10 [00:06<00:00,  1.58it/s]\n"
     ]
    }
   ],
   "source": [
    "denoising_and_sparse_scores = [get_scores(**config) for config in CONFIGS_DENOISING_AND_SPARSE]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 648x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_scores(CONFIGS_DENOISING_AND_SPARSE, denoising_and_sparse_scores, metric=\"auc\", savefile=\"plots/approach4/boxplot_kde_denoising_and_sparse.pdf\", figsize=(9, 8), label_below=[3,5,6], colors = [\"black\"] + [\"darkgoldenrod\"] * 5 + [\"blue\"] * 4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 648x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_scores(CONFIGS_DENOISING_AND_SPARSE, denoising_and_sparse_scores, metric=\"tnr95\", savefile=\"plots/approach4/boxplot_kde_denoising_and_sparse_tnr95.pdf\", figsize=(9, 8), label_below=[], colors = [\"black\"] + [\"darkgoldenrod\"] * 5 + [\"blue\"] * 4)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Compare sessions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Session 'Beaver_01' at folder: /home/kleinsteuber/vscode/ResizedSessions256_NoBackup/VIELAAS_Spring_Session01-VIELAAS_Beaver_01\n",
      "Loaded scans.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10/10 [00:02<00:00,  3.67it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Session 'Marten_01' at folder: /home/kleinsteuber/vscode/ResizedSessions256_NoBackup/VIELAAS_Spring_Session01-VIELAAS_Marten_01\n",
      "Loaded scans.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10/10 [00:13<00:00,  1.35s/it]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Session 'GFox_03' at folder: /home/kleinsteuber/vscode/ResizedSessions256_NoBackup/VIELAAS_Spring_Session03-VIELAAS_GFox_03\n",
      "Loaded scans.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10/10 [00:08<00:00,  1.18it/s]\n"
     ]
    }
   ],
   "source": [
    "compare_sessions_scores = [get_scores(**config) for config in CONFIGS_COMPARE_SESSIONS]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 360x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_scores(CONFIGS_COMPARE_SESSIONS, compare_sessions_scores, metric=\"auc\", savefile=\"plots/approach4/boxplot_kde_sessions.pdf\", figsize=(5, 7))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 360x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_scores(CONFIGS_COMPARE_SESSIONS, compare_sessions_scores, metric=\"tnr95\", savefile=\"plots/approach4/boxplot_kde_sessions_tnr95.pdf\", figsize=(5, 7))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.10.4 ('pytorch-gpu')",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.4"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "17cd5c528a3345b75540c61f907eece919c031d57a2ca1e5653325af249173c9"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}